Resumen de Intraurban Analysis of Surface Urban Heat Island From Disagregated Thermal Radiance Images
Vanessa da Silva Brum Bastos, Leila Maria Garcia Fonseca, Carolina Moutinho Duque de Pinho
Surface Urban Heat Islands (SUHI) are areas with higher surface temperatures than their surroundings. Several studies have used thermal images from satellites to research the influence of urbanization on surface temperature patterns, however the low spatial resolution of thermal sensors limits the analysis of LST intraurban variations. Attempting to overcome this limitation, we used the Enhanced Physical Model (EPM) for disaggregation of land surface temperature (DLST) to generate fine scale LST for Sao Paulo city in Brazil. This method uses a linear regression and Planck’s law to combine NDVI, NDWI and UI to estimate LST at finer spatial detail. First, we calibrate the method by upscaling an ASTER thermal band to 1000 m and using EPM to estimate the original 100 m thermal band. The original and estimated ASTER thermal bands achieved and R² of 0.66. Following, we apply the EPM model to estimate the LST at 15 m and compare it with data from meteorological stations. The 15 m LST image facilitated the identification of potential SUHIs. The EPM model provides an enhanced product with higher level of spatial detail, which allows researchers to identify changes of surface temperature that would not be evident from an ASTER LST (90 m spatial resolution) product. In summary, the model allowed us to quantify and map the influence of different urbanization patterns on the LST distribution.
